Re: Central Hurd web (Was: question about booting hurd)
On Tue, Dec 04, 2001 at 11:42:43AM +0100, Ondrej Hurt wrote:
> Look above for an idea of what I roughly mean by directory of Hurd
> specific stuff. I would merge it with hurd.gnu.org.
I like to compare the web site hurrd.gnu.org with the Hurd source code.
The Hurd source code is the essence of the Hurd development. It is what
is considered to be code worth to be included and maintained. We don't
include half-baken, incomplete, or broken code unless there is a strong
reason to do so. This actually has a real effect on the work needed to
maintain the code. If the first version you check in is near to perfect,
you don't need to make much changes later on.
Or, put the other way: If you just throw together some bits and pieces,
you can't hope that anything truly useful comes out of it. This holds
true for source code as well as for documentation and web sites.
A good link collection requires a lot of work and permanent maintenance.
It also requires a fine judgement about what to give how much weight.
Sure, all the details of current developments around the Hurd are obscured
by being spread all over the web. But google has failed me never to retrieve
any particular detail I lost. What is more devastating is that the very
basic, the documentation of the Hurd, its design, its tools and interfaces
is lacking. This is a serious problem. If someone visits hurd.gnu.org,
he wants to find out about the Hurd, and not about OSKit, L4, Microkernels
or any other subject which is related to the Hurd.
Don't get me wrong: at some time or another this information becomes important
for development. But for the first contact, it's all secondary.
> OK then the cetral site would just have a link to the document
> located somewhere else and the synch problems vanishes. You only have
> to check the state on linked documents.
This is an option, and we will certainly do so. A page with links to important
other sites will eventually come, for sure.
> I am really curious what sites is not GNU willing to link ?
GNU doesn't link to sites concerned about non-free software.
> But I really agree with you it is true that many people can get
> disappointed with the Hurd because of temporary problems (like I did
> year ago...). I would _strongly_ state at the GNU page that it is
> neither stable nor optimized from the performance point to view.
The "it's stable" hints at a technical design advantage that really exists,
so it is appropriate. Apart from that, it says nothing about practical
stability in day-to-day work. Maybe it should, suggest something.
It says that the Hurd is not ready for production use.
> Sure, the absolutely most appropriate place would be the GNU site but
> if you resist to link to some unwanted stuff and if you seem to be
> content with the current situation then I really don't know how to
> solve this...
Easy: Put your feet where your mouth is! Rather than
discussions about how things should or should not be, it's better
to just go forward and do something you think is a good idea, and then see
how it fits in.